Let $p$ be an odd prime number and let $\alpha = [X] \in R=\mathbb F_p[X]/\langle X^4+1\rangle$, and $y = \alpha + \alpha^{-1}$

I've proven:

1) $\alpha$ is a primitive eight root of unity in $R$.

2) $y^2 = 2$ and $y^p = \alpha^p + \alpha^{-p}$.

$\alpha^{-p} = {(\alpha^{-1})}^{p}$ right ?

3) $y^p = y$ if $p \equiv 1,7 \pmod 8$ and $y^p = -y$ if $p\equiv 3,5 \pmod 8$.

However I need help proving that $\left(\frac2p\right) = 1$ if $p \equiv 1,7 \pmod 8$ and $\left(\frac2p\right) = -1$ if $p \equiv 3,5 \pmod 8$. This should be accomplished using what I've already proven according to the text.

Please dont forget comment on my "$\alpha^{-p} = {(\alpha^{-1})}^{p}$ right ?".


  • $\begingroup$ I'd let $\pi(X)$ be a prime factor of $X^4+1$ and consider the ring $R=F_p[X]/\langle\pi(X)\rangle$, since this is a field, letting you do some simplifications. $\endgroup$ – Thomas Andrews Dec 5 '13 at 13:43
  • $\begingroup$ $R$ is indeed a field. Let $F$ be a field and $f \in F[X]$ an irreducible polynomial. Then $R=F[X]/\langle f \rangle$ is a field. $\endgroup$ – Shuzheng Dec 5 '13 at 13:43
  • $\begingroup$ No, $R$ is not necessarily a field. $X^4+1$ is not necessarily an irreducible polynomial. $\endgroup$ – Thomas Andrews Dec 5 '13 at 13:44
  • $\begingroup$ Ohh, guess you are right. But I really would like my answer to be based on 1,2,3. However since the problem assume $\alpha^{-1}$ exists doesn't that imply we are working in a field ? $\endgroup$ – Shuzheng Dec 5 '13 at 13:48

Yes, right, $\alpha^{-p} = \left(\alpha^{-1}\right)^p$.

Regarding the Legendre symbol, note that

$$y^p = \left(y^2\right)^{(p-1)/2}\cdot y = 2^{(p-1)/2}\cdot y = \left(\frac{2}{p}\right)\cdot y.$$

  • $\begingroup$ why $$2^{(p-1)/2}\cdot y = \left(\frac{2}{p}\right)\cdot y.$$ ? and could you tell me how we can assume $\alpha^{-1}$ exists, since we don't know if $R$ is a field. Thanks a lot btw. $\endgroup$ – Shuzheng Dec 5 '13 at 13:53
  • $\begingroup$ Is it because $$\left(\frac{2}{p}\right) \equiv 2^{(p-1)/2} (\mod p)$$ ? $\endgroup$ – Shuzheng Dec 5 '13 at 13:57
  • $\begingroup$ First $\alpha^{-1}$. The polynomial factored out is $X^4+1$, which means $\alpha^4 = -1$, and hence $\alpha^8 = 1$, so $\alpha^{-1} = \alpha^7$. Second, the Legendre symbol, yes, precisely, because $a^{(p-1)/2} \equiv \left(\frac{a}{p}\right) \pmod{p}$. $\endgroup$ – Daniel Fischer Dec 5 '13 at 13:59
  • $\begingroup$ I know $$\left(\frac{2}{p}\right) \equiv 2^{(p-1)/2}$$ but this just means that $\left(\frac{2}{p}\right) = 2^{(p-1)/2} + pk$. But then $$\left(\frac{2}{p}\right) \cdot y = (2^{(p-1)/2} + pk) \cdot y = 2^{(p-1)/2} \cdot y$$ why ? $\endgroup$ – Shuzheng Dec 5 '13 at 14:28
  • $\begingroup$ @user111854 We're working over $\mathbb{F}_p$, so all integers, like $2^{(p-1)/2}$ or $\left(\frac{2}{p}\right)$ are reduced modulo $p$/represent their residue class modulo $p$. In $\mathbb{F}_p$, we have the equality, $a^{(p-1)/2} = \left(\frac{a}{p}\right)$, not only the congruence. $\endgroup$ – Daniel Fischer Dec 5 '13 at 14:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.